Profiling adaptive immune repertoires across multiple human tissues by RNA Sequencing

In a project led by Serghei Mangul, members of our lab recently developed and tested a novel computational method that uses regular RNA-Seq data to rapidly and accurately profile the human immune system. Mangul and his collaborators, including UCLA graduate student Harry (Taegyun) Yang and 2016 B. I. G. Summer undergraduate participants Jeremy Rotman, Benjamin Statz, and Will Van Der Wey, recently published their results in a paper on bioRxiv.

Discoveries in human immunology and advancements in development of treatments for many common human diseases depend on detailed reconstructions of the adaptive immune repertoire. The “adaptive” immune repertoire recognizes pathogens and toxins that the “innate” defense system misses. Assay-based genetic studies provide a detailed view of these adaptive systems by profiling the genetic expression and repertoires of B and T cell receptors. Assay-based approaches have accurately characterized the immune repertoire of peripheral blood.

However, these methods are expensive and smaller in scale when compared to standard RNA sequencing (RNA-seq). Characterizing the immunological repertoires of other tissues, including barrier tissues like skin and mucosae, requires large-scale study. RNA-Seq can capture the entire cellular population of a sample, including B and T cell and their receptors.

ImReP is the first method to efficiently extract B and T cell receptor derived reads from RNA-Seq data, accurately assemble CDR3 sequences, the most variable regions of these receptors, and determine their antigen specificity. Mangul and his team used simulated data to test the feasibility of using RNA-Seq to study the adaptive immune repertoire. ImReP is able to identify 99% CDR3-derived reads from the RNA-Seq mixture, suggesting it is a powerful tool for profiling RNA-Seq samples of immune-related tissues.

They also compared methods and investigated the sequencing depth and read length required to reliably assemble B and T cell receptor sequences from RNA-Seq data. ImReP consistently outperformed existing methods in both recall and precision rates for the majority of simulated parameters. Notably, ImReP was the only method with acceptable performance at 50bp read length, reconstructing with higher precision rate significantly more CDR3 clonotypes.

Mangul and his team applied ImReP to 8,555 samples across 544 individuals from 53 tissues obtained from Genotype-Tissue Expression study (GTEx v6). The data was derived from 38 solid organ tissues, 11 brain subregions, whole blood, and three cell lines. ImRep identified over 26 million reads overlapping 3.8 million distinct CDR3 sequences that originate from diverse human tissues.

Using ImReP, they created a systematic atlas of immunological sequences for B and T cell repertoires across a broad range of tissue types, most of which were not previously studied for B and T cell repertoires. They also examined the compositional similarities of clonal populations between tissues to track the flow of B and T clonotypes across immune-related tissues, including secondary lymphoid and organs encompassing mucosal, exocrine, and endocrine sites.

Advantages of using RNA-Seq to study immune repertoires include the ability to simultaneously capture both B and T cell clonotype populations during a single run, simultaneously detect overall transcriptional responses of the adaptive immune system, and scaling up the atlas of B and T cell receptors that will provide valuable insights into immune responses across various autoimmune diseases, allergies, and cancers.

Read more about ImReP in the full article, which is available for download on bioRxiv

ImReP was created by Igor Mandric and Serghei Mangul. ImReP is freely available at:

The atlas of T and B cell receptors, the largest collection of CDR3 sequences and tissue types, is freely available at This resource has potential to enhance future studies in areas such as immunology and advance development of therapies for human diseases.

The full citation to our paper is:

Mangul, S., Mandric, I., Yang, H.T., Strauli, N., Montoya, D., Rotman, J., Van Der Wey, W., Ronas, J.R., Statz, B., Zelikovsky, A. and Spreafico, R., 2016. Profiling adaptive immune repertoires across multiple human tissues by RNA Sequencing. bioRxiv, p.089235.


Figure 1. Overview of ImReP.

Figure 1. Overview of ImReP. (See full paper for details.)


Figure 6. Flow of T and B cell clonotypes across diverse human tissues.

Figure 6. Flow of T and B cell clonotypes across diverse human tissues. (See full paper for details.)


Colocalization of GWAS and eQTL Signals Detects Target Genes

Farhad Hormozdiari recently developed a method for combining genome-wide association studies (GWASs) and quantitative trait loci (eQTL) studies in a statistical framework that quantifies the probability of each variant to be causal while allowing an arbitrary number of causal variants. Together with collaborators at the University of Oxford and Broad Institute of MIT and Harvard, we present a paper in The American Journal of Human Genetics. Here, we describe eQTL and GWAS CAusal Variants Identification in Associated Regions (eCAVIAR). We apply our approach to datasets from several GWASs and eQTL studies in order to assess its accuracy and potential contributions to colocalization and fine-mapping.

Integrating GWASs and eQTL studies is a promising way to explore the mechanism of non-coding variants on diseases. Integration of GWAS and eQTL data is challenging due to the uncertainty induced by linkage disequilibrium (LD), the non-random association of alleles at different loci, and presence of loci that harbor multiple causal variants (allelic heterogeneity). Current methods assume that each locus contains a single causal variant and expect loci to be independent and associated randomly.

eCAVIAR is a novel probabilistic model for integrating GWAS and eQTL data that extends the CAVIAR (Hormozdiari et al. 2014) framework to explicitly estimate the posterior probability of the same variant being causal in both GWAS and eQTL studies, while accounting for allelic heterogeneity and LD. Our approach can quantify the strength between a causal variant and its associated signals in both studies, and it can be used to colocalize variants that pass the genome-wide significance threshold in GWAS. For any given peak variant identified in GWAS, eCAVIAR considers a collection of variants around that peak variant as one single locus.

We apply eCAVIAR to the Meta-Analyses of Glucose and Insulin-related traits Consortium (MAGIC) dataset and GTEx dataset to detect the target gene and most relevant tissue for each GWAS risk locus. When applied to the MAGIC dataset’s 2 phenotypes, eCAVIAR identifies genetic variants that are causal in both eQTL and GWAS. Further, eCAVIAR detects a large number of loci where the GWAS causal variants are clearly distinct from the causal variants in the eQTL data. Interestingly, eCAVIAR also identifies genes that colocalize in one tissue yet can be excluded in others. For the majority of loci in which we identify a single variant causal for both GWAS and eQTL, eCAVIAR implicates more than one causal variant across the 45 tissues.

We observe that eCAVIAR outperforms existing methods even when there are different values of non-colocalization. Using simulated datasets, we compared accuracy, precision, and recall rate of eCAVIAR to RTC (Nica et al. 2010) and COLOC (Giambartolomei et al. 2014), two current methods for eQTL and GWAS colocalization. Our results show that eCAVIAR has high confidence for selecting loci to be colocalized between the GWAS and eQTL data and is conservative in selecting a locus to be colocalized.

We hope that future applications of eCAVIAR will advance identification of specific GWAS loci that share a causal variant with eQTL studies in a tissue, thus providing insight into presently unclear disease mechanisms.


Overview of eCAVIAR.


eCAVIAR was created by Farhad Hormozdiari, Ayellet V. Segre, Martijn van de Bunt, Xiao Li, Jong Wha J Joo, Michael Bilow, Jae Hoon Sul, Bogdan Pasaniuc and Eleazar Eskin. The article is available at:

Visit the following page to download CAVIAR and eCAVIAR:

The full citation to our paper is:

Hormozdiari, Farhad; van de Bunt, Martijn; Segrè, Ayellet V; Li, Xiao; Joo, Jong Wha J; Bilow, Michael; Sul, Jae Hoon; Sankararaman, Sriram; Pasaniuc, Bogdan; Eskin, Eleazar

Colocalization of GWAS and eQTL Signals Detects Target Genes. Journal Article

In: Am J Hum Genet, 2016, ISSN: 1537-6605.

Abstract | Links | BibTeX

Our paper builds upon a method introduced in a previous publication:

Hormozdiari, Farhad; Kostem, Emrah ; Kang, Eun Yong ; Pasaniuc, Bogdan ; Eskin, Eleazar

Identifying causal variants at Loci with multiple signals of association. Journal Article

In: Genetics, 198 (2), pp. 497-508, 2014, ISSN: 1943-2631.

Abstract | Links | BibTeX

Discovering SNPs Regulating Human Gene Expression Using Allele Specific Expression from RNA-Seq data

Analyses of expression quantitative trait loci (eQTL), genomic loci that contribute to variation in genetic expression levels, are essential to understanding the mechanisms of human disease. These studies identify regulators of gene expression as either cis-acting factors that regulate nearby genes, or trans-acting factors that affect unlinked genes through various functions.  Traditional eQTL studies treat expression as a quantitative trait and associate it with genetic variation. This approach has identified many loci involved in the genetic regulation of common, complex diseases.

Standard eQTL methods are limited in power and accuracy by several phenomena common to genomic datasets. First, the correlation structure of genetic variation in the genome, known as linkage disequilibrium (LD), limits the ability of these methods to differentiate between the regulatory variant and neighboring variants that are in LD. Second, like other quantitative traits, the total expression of a gene is influenced by multiple genetic and environmental factors. The effect size for any given variant is therefore small, and standard methods require a large sample size to identify the effect.


ASE example and corresponding mathematical representation of three individuals (1, 2, 3). We assume that the third SNP is the causal SNP site affecting the differential gene expression level (Allele A/ Allele T).

Our forthcoming paper in Genetics presents a new method that improves the accuracy and computational power of eQTL mapping with incorporation of allele specific expression (ASE) analysis. Our novel method uses genome sequencing, alongside measurements of ASE from RNA-seq data, to identify cis-acting regulatory variants.

In standard eQTLs studies, the analysis of ASE is influenced by LD structure and the amount of allelic heterogeneity present in the genome. Individual effects appear weak since the effect of a variant is modest when compared to the variance of total expression. In our approach, the genotypes of each single individual with ASE provides information useful to determining variants causal for the observed ASE. Our approach actually leverages the relationship between LD and variant identification to map the variants affecting expression. Thus, analysis of ASE is advantageous over analysis of total expression levels, the standard approach to eQTL mapping.

We demonstrate the utility of our method by analyzing RNA-seq data from 77 unrelated northern and western European individuals (CEU). To map each gene, we simultaneously compare ASE measurements across a set of sequenced individuals. We then identify genetic variants that are in proximity to those genes and capable of explaining observed patterns of ASE. Here, we characterize the efficacy of this method as the ratio termed “reduction rate” and denoted as the ratio between the number of candidate regulatory SNPs to the total number of SNPs in the proximal region of the gene.

When applied to the CEU dataset, our method reduced the set of candidate SNPs from ten to two (a reduction rate of 80%). Allowing for one error increases the number of candidate SNPs to five and decreases the reduction rate to 50%. We also observe that the relationship between LD and variant identification has a different quality in ASE mapping when compared to eQTL studies, and produces different types of information useful to eQTL mapping studies.

ASE studies are a powerful approach to identifying associations between genetic variation and gene expression. Accurate measurement of ASE can identify cis-acting regulatory variants associated with common diseases. Our novel method for ASE mapping is based on a robust and computationally efficient non-parametric approach, and we hope it advances our understanding of functional risk alleles and facilitates development of new hypotheses for the causes and treatment of common diseases.

This project used software developed by Jennifer Zou, which is available for download at:

This project was led by Eun Yong Kang and involved Serghei Mangul, Buhm Han, and Sagiv Shifman. The article is available at:

The full citation to our paper is:

Kang, Eun Yong; Martin, Lisa; Mangul, Serghei; Isvilanonda, Warin; Zou, Jennifer; Ben-David, Eyal; Han, Buhm; Lusis, Aldons J; Shifman, Sagiv; Eskin, Eleazar

Discovering SNPs Regulating Human Gene Expression Using Allele Specific Expression from RNA-Seq Data. Journal Article

In: Genetics, 2016, ISSN: 1943-2631.

Abstract | Links | BibTeX